Agent Development
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection Surface: The skill provides templates in
SKILL.mdandexamples/agent-creation-prompt.mdthat facilitate AI-assisted agent generation by interpolating user-supplied descriptions into a system prompt. This creates a surface where untrusted data could influence the generated agent's instructions. - Ingestion points: The
[YOUR DESCRIPTION]placeholder inSKILL.mdand the usage patterns inexamples/agent-creation-prompt.mdare primary entry points for untrusted data. - Boundary markers: The prompt templates currently lack explicit boundary markers or delimiters to isolate user-supplied descriptions from the agent's meta-instructions.
- Capability inventory: Generated agents are shown to have access to powerful tools including
Bash,Write,Read,Glob, andGrepin the provided examples (examples/complete-agent-examples.md). - Sanitization: There is no evidence of sanitization or filtering of the user-supplied description before it is processed by the agent architect prompt.
- Security Validation Script: The
scripts/validate-agent.shutility is a benign bash script designed to verify the structural integrity, naming conventions, and frontmatter of agent files. It uses standard text processing tools (grep,sed,awk) and does not perform network operations or privilege escalation. - Security Best Practices: The skill proactively includes a
security-analyzer.mdagent template in its examples. This template instructs agents to look for vulnerabilities such as SQL injection, XSS, and hardcoded secrets, which promotes secure coding practices within the agent development ecosystem.
Audit Metadata